Social Categorization and Stereotyping: How Social Categorization Biases Person Perception and Face Memory

The current article offers a brief history of research and theory related to social categorization and stereotyping. We begin by outlining research focusing on category selection, category activation, and category application. We then discuss the consequences of social categorization, noting that one of the most important and ubiquitous effects of social categorization is homogenization or assimilation. We then extend this research related to the assimilative consequences of social categorization to the realm of face memory. Specifically, we introduce the Categorization‐Individuation Model (see also Hugenberg et al., forthcoming) as a new model of the Cross‐Race Effect, or the difficulty many perceivers have in recognizing members of other racial groups. The Categorization‐Individuation Model argues that the Cross‐Race Effect is due to the tendency to categorize out‐group members but individuate in‐group members. Finally, we summarize the extant research in support of this new model of person memory and close by discussing the current state of, and potential future directions for, social categorization research.     

群体交叉分类效应的代数模型及其潜在加工过程

交叉分类被认为是一种能够有效减少群际偏见的方法。交叉分类效应的代数模型主要包括平均模型、相加模型和等价模型,各个模型都有自己成立的证据,其中支持相加模型的研究证据最多。尽管如此,交叉分类效应的代数模型有很明显的局限,即要求所有进行交叉的类别维度权重相等,这与真实社会情境是相矛盾的。交叉分类的潜在加工过程主要有基于认知角度的类别差异模型和基于动机角度的社会认同两种比较成熟的理论,前者又分为类别差异消除模型和类别差异减少模型。     

A generalized longitudinal mixture IRT model for measuring differential growth in learning environments

This article describes a generalized longitudinal mixture item response theory (IRT) model that allows for detecting latent group differences in item response data obtained from electronic learning (e-learning) environments or other learning environments that result in large numbers of items. The described model can be viewed as a combination of a longitudinal Rasch model, a mixture Rasch model, and a random-item IRT model, and it includes some features of the explanatory IRT modeling framework. The model assumes the possible presence of latent classes in item response patterns, due to initial person-level differences before learning takes place, to latent class-specific learning trajectories, or to a combination of both. Moreover, it allows for differential item functioning over the classes. A Bayesian model estimation procedure is described, and the results of a simulation study are presented that indicate that the parameters are recovered well, particularly for conditions with large item sample sizes. The model is also illustrated with an empirical sample data set from a Web-based e-learning environment.     

Automaticity in rule-based and information-integration categorization

Three experiments studied the effects of category structure on the development of categorization automaticity. In Experiment 1, participants were each trained for over 10,000 trials in a simple categorization task with one of three category structures. Results showed that after the first few sessions, there were no significant behavioral differences between participants who learned rule-based versus information-integration category structures. Experiment 2 showed that switching the locations of the response keys after automaticity had developed caused a similar highly significant interference, regardless of category structure. In Experiment 3, a simultaneous dual task that engaged executive functions did not interfere with either rule-based or information-integration categorization. These novel results are consistent with a theory assuming separate processing pathways for initial rule-based and information-integration category learning but a common processing pathway after the development of automaticity.     

Domain differences in absolute judgments of category membership: evidence for an essentialist account of categorization

There has been some debate about the correspondence between typicality gradients and category membership. The present study investigates the relationship between these two measures in the domains of animals and artifacts. Forty-two adults judged the degree of typicality or category membership of 293 animals and artifacts. The subjects’ tendency for animals, but not for artifacts, was to make more absolute ratings on category membership (i.e., judging exemplars as definitely members or definitely not members of their respective category) than on typicality. More importantly, at almost every level of typicality, subjects were more likely to make absolute judgments of category membership for animals than for artifacts. These results indicate that people treat category membership of animals as relatively absolute (which best fits an essentialist model of categorization) and treat category membership of artifacts as relatively graded (which best fits a prototype model of categorization). These domain differences add crucial supporting evidence for claims about the domain-specificity of essentialism.     

Using the open-source statistical language R to analyze the dichotomous Rasch model

R, an open-source statistical language and data analysis tool, is gaining popularity among psychologists currently teaching statistics. R is especially suitable for teaching advanced topics, such as fitting the dichotomous Rasch model-a topic that involves transforming complicated mathematical formulas into statistical computations. This article describes R’s use as a teaching tool and a data analysis software program in the analysis of the Rasch model in item response theory. It also explains the theory behind, as well as an educator’s goals for, fitting the Rasch model with joint maximum likelihood estimation. This article also summarizes the R syntax for parameter estimation and the calculation of fit statistics. The results produced by R is compared with the results obtained from MINI STEP and the output of a conditional logit model. The use of R is encouraged because it is free, supported by a network of peer researchers, and covers both basic and advanced topics in statistics frequently used by psychologists.     

Category width and differentiation in semantic categories

The hypothesis that the judgemental categorizing behaviour measured by Pettigrew's Category-Width scale is linked to semantic categorization is investigated. Individual differences in broad vs. narrow categorizing behaviour measured by the C-W scale are reconceptualized within the framework of Rosch's (1977) theory of human categorization. Two extreme groups on the C-W scale, broad categorizers and narrow categorizers, were selected as subjects. One experiment employed a short-term release-from-proactive-inhibition paradigm. The amount of discrepancy in recall between the preshift and the shift trials was used as an index to perceived semantic distance. The result showed that narrow categorizers perceived the same fuzzy items in a semantic category as less similar to the good exemplars of that category than broad categorizers did. In a further experiment using a semantic-categorization technique, narrow categorizers judged the same fuzzy items as category members less frequently than broad categorizers did, indicating that narrow categorizers possess more differentiated semantic categories than do broad categorizers. These findings are seen as an extension of the differential categorizing behaviour on artificial categories in the C-W scale to semantic categorization for which there is a pre-existing system.     

Attention, similarity, and the identification-categorization relationship

A unified quantitative approach to modeling subjects' identification and categorization of multidimensional perceptual stimuli is proposed and tested. Two subjects identified and categorized the same set of perceptually confusable stimuli varying on separable dimensions. The identification data were modeled using Shepard's (1957) multidimensional scaling-choice framework. This framework was then extended to model the subjects' categorization performance. The categorization model, which generalizes the context theory of classification developed by Medin and Schaffer (1978), assumes that subjects store category exemplars in memory. Classification decisions are based on the similarity of stimuli to the stored exemplars. It is assumed that the same multidimensional perceptual representation underlies performance in both the identification and categorization paradigms. However, because of the influence of selective attention, similarity relationships change systematically across the two paradigms. Some support was gained for the hypothesis that subjects distribute attention among component dimensions so as to optimize categorization performance. Evidence was also obtained that subjects may have augmented their category representations with inferred exemplars. Implications of the results for theories of multidimensional scaling and categorization are discussed.     

Monte Carlo tests of the Rasch model based on scalability coefficients

For item responses fitting the Rasch model, the assumptions underlying the Mokken model of double monotonicity are met. This makes non‐parametric item response theory a natural starting‐point for Rasch item analysis. This paper studies scalability coefficients based on Loevinger's H coefficient that summarizes the number of Guttman errors in the data matrix. These coefficients are shown to yield efficient tests of the Rasch model using p‐values computed using Markov chain Monte Carlo methods. The power of the tests of unequal item discrimination, and their ability to distinguish between local dependence and unequal item discrimination, are discussed. The methods are illustrated and motivated using a simulation study and a real data example.     

In search of abstraction: The varying abstraction model of categorization

A longstanding debate in the categorization literature concerns representational abstraction. Generally, when exemplar models, which assume no abstraction, have been contrasted with prototype models, which assume total abstraction, the former models have been found to be superior to the latter. Although these findings may rule out the idea that total abstraction takes place during category learning and instead suggest that no abstraction is involved, the idea of abstraction retains considerable intuitive appeal. In this article, we propose the varying abstraction model of categorization (VAM), which investigates the possibility that partial abstraction may play a role in category learning. We apply the VAM to four previously published data sets that have been used to argue that no abstraction is involved. Contrary to the previous findings, our results provide support for the idea that some form of partial abstraction can be used in people's category representations.     

Loglinear Rasch model tests

Existing statistical tests for the fit of the Rasch model have been criticized, because they are only sensitive to specific violations of its assumptions. Contingency table methods using loglinear models have been used to test various psychometric models. In this paper, the assumptions of the Rasch model are discussed and the Rasch model is reformulated as a quasi-independence model. The model is a quasi-loglinear model for the incomplete subgroup × score × item 1 × item 2 × ... × itemk contingency table. Using ordinary contingency table methods the Rasch model can be tested generally or against less restrictive quasi-loglinear models to investigate specific violations of its assumptions.     

Contextual changes in the prototypicality of extreme and moderate outgroup members

Previous social categorization research has tended to treat prototypicality (the degree to which a stimulus is representative of a category) as a fixed stimulus property. In contrast, self-categorization theory sees prototypicality as an aspect of the categorization process that is dependent on features of the social context within which categorization takes place. To test this view two experiments (Ns = 256, 73) examined the perceived representativeness of extreme and moderate members of the same target outgroup in conditions which manipulated the salience of intergroup division. As predicted, the extremist was seen to be relatively more representative of the outgroup than the moderate to the extent that intergroup differences were salient. In Experiment 1 the extremist's message was also seen to be less important in low salience conditions and in Experiment 2 shifts in prototypicality were associated with changes in the evaluation of individual targets. Implications for the analysis of social categorization and stereotyping are discussed.     

Thirty categorization results in search of a model

One category structure dominated in the shift toward exemplar-based theories of categorization. Given the theoretical burden on this category structure, the authors reanalyzed 30 of its uses over 20 years in 8 articles. The authors suggest 4 conclusions. (a) This category structure may encourage exemplar-memorization processes because of its poor structure, the learning difficulties it causes, and its small, memorizable exemplar sets. Its results may only generalize narrowly. (b) Exemplar models have an advantage in fitting these 30 data sets only because they reproduce a performance advantage for training items. Other models fit equally well if granted this capacity. (c) A simpler exemplar process than assumed by exemplar models suffices to explain these data sets. (d) An important qualitative result predicted by exemplar theory is not found overall and possibly should not even be expected. The authors conclude that the data produced by this category structure do not clearly support exemplar theory.     

Instance memorization and category influence: Challenging the evidence for multiple systems in category learning

A class of dual-system theories of categorization assumes a categorization system based on actively formed prototypes in addition to a separate instance memory system. It has been suggested that, because they have used poorly differentiated category structures (such as the influential “5-4” structure), studies supporting the alternative exemplar theory reveal little about the properties of the categorization system. Dual-system theories assume that the instance memory system only influences categorization behaviour via similarity to single isolated instances, without generalization across instances. However, we present the results of two experiments employing the 5-4 structure to argue against this. Experiment 1 contrasted learning in the standard 5-4 structure with learning in an even more poorly differentiated 5-4 structure. In Experiment 2, participants memorized the 5-4 structure based on a five minute simultaneous presentation of all nine category instances. Both experiments revealed category influences as reflected by differences in instance learnability and generalization, at variance with the dual-system prediction. These results have implications for the exemplars versus prototypes debate and the nature of human categorization mechanisms.     

Categorization by groups and individuals

Categorization is a core psychological process that is central to decision making. While a substantial amount of research has been conducted to examine individual categorization behavior, little is known about how the outputs of individual and group categorization may differ. Four experiments demonstrate that group categorization differs systematically from individual categorization in the structural dimension of category breadth: categorizing the same set of items, groups tend to create a larger number of smaller categories than individuals. This effect of social context is a function of both taskwork and teamwork. In terms of taskwork, groups’ greater available knowledge mediates differences in category breadth between individuals and groups by increasing utilized knowledge (study 2). In terms of teamwork, task conflict moderates the effect of social context on category breadth (study 3). Moreover, the experience of categorizing individually or in a group influences individuals’ subsequent judgments (study 4).     

Analyzing experimental data using the Rasch model

We present a method for studying experimental data based on a psychometric model, the “Rasch model” (Rasch, 1966; Thissen & Steinberg, 1986). We illustrate the method with the use of a data set in the field of concept research. More specifically, we investigate whether a conjunctive concept can be seen as an additive combination of its constituents. High correlations between model and data are obtained, but a formal goodness-of-fit test indicates that the model does not completely account for the data. We then alter the Rasch model in such a way as to capture our idea of why the model deviates from the data. This results in higher correlations and a strong increase in goodness-of-fit. It is concluded that our ideas, as incorporated in the model, adequately summarize the data. More generally, this research illustrates that applying the Rasch model and altering it according to one’s hypotheses is an excellent way to analyze experimental data.     

The rasch model,the law of comparative judgment and additive conjoint measurement

Relationships between the Rasch model and both the law of comparative judgment and additive conjoint measurement are discussed. The distance between the ability of Persona and the difficult of Itemi is, in the Rasch model, the baseline value corresponding to the probability thata will respond correctly toi, where this probability is interpreted as the area under a logistic curve (which is substantially equivalent to the normal curve) and is thus an application of the law of comparative judgment. Under certain assumptions, the Rasch model is also a special case of additive conjoint measurement and, properly reinterpreted, may be usefully applied in contexts other than individual differences.     

Matching words to concepts in two languages: A test of the concept mediation model of bilingual representation

A categorization paradigm was used to investigate the relations between lexical and conceptual connections in bilingual memory. Fifty-one more fluent and less fluent English-French bilinguals viewed category names (e.g.,vegetable) and then decided whether a target word (e.g.,peas) was a member of that category. The category names and target words appeared in both English and French across experimental conditions. Because category matching requires access to conceptual memory, only relatively fluent bilinguals, who are able to directly access meaning for their second language, were expected to be able to use category information across languages. The performance of lessfluent bilinguals was expected to reflect reliance on lexical-level connections between languages, requiring translation of second-language words. The results provided evidence for concept mediation by more-fluent bilinguals, because categorization latencies were independent of the language of the category name. However, the performance of less-fluent bilinguals indicated that they did not follow a simple lexical translation strategy. Instead, these subjects were faster at categorizing words in both languages when the language of the category name matched the language of the target word, suggesting that they were able to access limited conceptual information from the second language. A model of the development of concept mediation during second language acquisition is described.     

A relation between a between-item multidimensional IRT model and the mixture rasch model

Two generalizations of the Rasch model are compared: the between-item multidimensional model (Adams, Wilson, and Wang, 1997), and the mixture Rasch model (Mislevy & Verhelst, 1990; Rost, 1990). It is shown that the between-item multidimensional model is formally equivalent with a continuous mixture of Rasch models for which, within each class of the mixture, the item parameters are equal to the item parameters of the multidimensional model up to a shift parameter that is specific for the dimension an item belongs to in the multidimensional model. In a simulation study, the relation between both types of models also holds when the number of classes of the mixture is as small as two. The relation is illustrated with a study on verbal aggression. Frank Rijmen was supported by the Fund for Scientific Research Flanders (FWO). This research is also funded by the GOA/2000/02 granted from the KU Leuven. We would like to thank Kristof Vansteelandt for providing the data of the study on verbal aggression.     

Domain differences in the structure of artifactual and natural categories

In three experiments, different methodologies, measures, and items were employed to address the question of whether, and to what extent, membership in a semantic category is all or none (i.e., absolute) or a matter of degree (i.e., graded). Resemblance theory claims that categorization is based on similarity, and because similarity is graded, category membership may also be graded. Psychological essentialism asserts that categorization is based on the presumption of the category essence. Because artifactual (e.g., FURNITURE) and natural (e.g., FRUIT) categories have different sorts of essences, artifacts and natural kinds may be categorized in qualitatively different manners. The results converged onthe finding of a robust domain difference in category structure: Artifactual categories were more graded than natural categories. Furthermore, typicality reliably predicted absolute category membership, but failed to predict graded category membership. These results suggest that resemblance theory and psychological essentialism may provide a concerted account of representation and categorization across domains.